The Department of Commerce (DoC) established a new data working group this week that aims to prepare public data to be AI-ready, according to DoC Chief Data Officer (CDO) Oliver Wise.
This group will explore more formally what it means for data to be AI-ready, but Wise explained that the group’s current working approach to that is “preparing data that’s not just machine-readable, but machine-interpretable.”
“What it comes down to is publishing data in a way that’s tidy data, if you’re in that world, where each column is a variable, each row is an observation,” Wise said today at the Data and Analytics Summit hosted by Nextgov/FCW in Washington. “So, just well-organized data, one, and two, that that data is the standard schema and has sufficient metadata for a crawler to understand.”
Wise, who chairs the Commerce Data Governance Board (CDGB), said the idea of the new working group came to fruition from a conversation about how the CDGB can leverage AI to advance its data democratization goals.
The working group will have two core deliverables, according to Wise. One, he said is to articulate guidelines or technical requirements for publishing AI-ready data.
“We’ve been doing this kind of in a nimble, informal way for the last year or so,” Wise said. “Our counterparts across the world that are in Europe are doing similar efforts, but we hope to enshrine that in something more formal and have an engagement process around that.”
The second core deliverable, he said, is to develop “a how-to guide” on the best practices for publishing that data.
“Basically, you’re going to define the goalposts, like what are the requirements, and then what are the smart effective ways to achieve those – kick the goal in the goalpost.”
The CDO added that while “it’s not entirely clear what the standards should be” when it comes to AI, the DoC is hoping that this working group can contribute to that conversation, “and then we also just hope to share tactical best practices on how to get there.”
Wise said DoC plans to collaborate with other Federal agencies, as well as international partners who are looking at the same issue of “how do we develop an ecosystem where users can use natural language interfaces when exploring data?”
The working group will be made up of data management leads. The group’s chair is Sallie Ann Keller, the chief scientist and associate director of the U.S. Census Bureau’s Research and Methodology Directorate, Wise said.